AI Agent Hub
Back to models
tiny_starcoder_py logo

tiny_starcoder_py

Code bigcode Released 2023-05-15
-- 0.164B params 8.2K context Open Source

About this model

TinyStarCoderPy

This is a 164M parameters model with the same architecture as StarCoder (8k context length, MQA & FIM). It was trained on the Python data from StarCoderData for ~6 epochs which amounts to 100B tokens.

Use

Intended use

The model was trained on GitHub code, to assist with some tasks like Assisted Generation. For pure code completion, we advise using our 15B models StarCoder or StarCoderBase.

Generation

# pip install -q transformers
from transformers import AutoModelForCausalLM, AutoTokenizer

checkpoint = "bigcode/tiny_starcoder_py"
device = "cuda" # for GPU usage or "cpu" for CPU usage

tokenizer = AutoTokenizer.from_pretrained(checkpoint)
model = AutoModelForCausalLM.from_pretrained(checkpoint).to(device)

inputs = tokenizer.encode("def print_hello_world():", return_tensors="pt").to(device)
outputs = model.generate(inputs)
print(tokenizer.decode(outputs[0]))

Fill-in-the-middle

Fill-in-the-middle uses special tokens to identify the prefix/middle/suffix part of the input and output:

input_text = "<fim_prefix>def print_one_two_three():\n    print('one')\n    <fim_suffix>\n    print('three')<fim_middle>"
inputs = tokenizer.encode(input_text, return_tensors="pt").to(device)
outputs = model.generate(inputs)
print(tokenizer.decode(outputs[0]))

Training

Model

  • Architecture: GPT-2 model with multi-query attention and Fill-in-the-Middle objective
  • Pretraining steps: 50k
  • Pretraining tokens: 100 billion
  • Precision: bfloat16

Hardware

  • GPUs: 32 Tesla A100
  • Training time: 18 hours

Software

License

The model is licensed under the BigCode OpenRAIL-M v1 license agreement. You can find the full agreement here.

Benchmark Scores

HumanEval
7.84

Technical Specs

  • Parameters: 0.164B
  • Architecture: GPT-2
  • Context Window: 8,192 tokens
  • Input Modalities: text

Hardware Requirements

  • VRAM: 1.0 GB
  • Compute: Runs on CPU or GPU with ~1 GB VRAM for inference; trained on 32× Tesla A100 GPUs using Megatron-LM